The goal of Project STARRT was to increase adolescents' understanding of AIDS risks and test a preventive intervention to decrease risk behaviors for pregnancy, STDs and AIDS. A social skills training program was developed to impart peer-resistance skills for drug and sexual practices among 440 Latino and Anglo youth, and to determine the degree of dissemination of skills to friends. The random assignment design and objective measures of skills made it possible to assess the role of social network interactions on AIDS risk practices. The experimental training resulted in a four-fold increase in social skills levels, and evidence of dissemination of skills between trained youth and their friends was observed. In addition, correlates of social skills and risk behaviors supported our conceptual model of adolescent sexual behavior. This competitive renewal proposes to follow these youth through four years post training. At which time they will be 18 to 23 years of age. The majority will be sexually active and many will have over 5 years of sexual experience. This follow-up will require 18 months of continued support and will enable the assessment of the durability of training effects as well as exploration of the social determinants of risk practices. Planned analysis will contribute to the understanding of AIDS risk during the period of life when approximately 20% of all HIV infections are established, and for under-served populations in the case of Latinos. The proposed competitive continuation of Project STARRT will take advantage of existing reliable measures, trained staff, data collection facilities and a comprehensive data set on a cohort demonstrated as retainable for additional measures. These pre-existing factors make such an extension an efficient means of collecting longitudinal data from early adolescence through early adulthood for Latino and Anglo youth. The following specific aims will be achieved during the proposed 18-month continuation period: 1. to recruit at least 85% of the initial cohort to a 48-month follow-up interview of risk behavior and disease status; 2. to determine the differential effects of the social skills training, usual sex education and no-training on risk practices and outcomes at 48 months; 3. to explore the relationship between changes in hypothesized determinants of sexual behavior and changes in sexual behaviors.